Ofcom Consultation: Promoting investment and competition in fibre networks
UKCTA Response to Ofcom
1. UKCTA is a trade association promoting the interests of fixed-line telecommunications companies competing against BT, as well as each other, in the residential and business markets. We advocate regulatory outcomes designed to serve consumer interests, particularly through competition to Ofcom and the Government. Details of membership of UKCTA can be found at www.ukcta.org.uk.
2. In the event of a conflict in the views between this submission and the submissions of individual UKCTA members, individual submissions will take priority. The views set out in this response do not necessarily reflect those of Virgin Media. UKCTA members welcome the opportunity to comment on Ofcom’s proposals to promote fibre investment.
3. In the remainder of this submission we set out our views on:
a. The task of co-ordinating the review of multiple markets to ensure consistency, whilst ensuring the remedies are appropriate to address SMP;
b. The importance of the Product Market Definition as a starting point;
c. Geographic Market Definition;
Co-ordination requirements for Multiple Markets
4. UKCTA members generally agree with conducting market analyses of physical infrastructure, local access, business connectivity, and broadband access at the same time.
5. UKCTA notes that such an approach has some advantages, including:
a. using the same data and financial information for multiple reviews limits discrepancies between reviews;
b. issues which are common to different markets can be dealt with and considered holistically (reducing “parking” and the forgetting of issues between reviews);
c. ensures consistency of remedies across markets; and
d. reduces the risk of manipulation or cost shifting by Openreach between reviews.
6. However, UKCTA members note that there are also risks to this approach which Ofcom must consider and address, including:
a. The combined reviews will represent a large exercise for Ofcom and industry and will require significant resource on both sides. Ofcom must ensure it has this available, and that this resource is fully up to speed on both residential vs business markets and all relevant technologies in consideration. Ofcom must also consider the impact on industry stakeholders and be flexible in its working relationship, for example around information gathering.
b. Larger market reviews and decisions could increase the impact of any legal challenge since a single successful appeal will result in a very wide regulatory lacuna, likely spanning multiple markets. UKCTA would not want to see the uncertainty of the BCMR Temporary Conditions situation happen for the large swathes of the relevant access markets. Ofcom must manage this risk and provide assurances to industry.
c. As Ofcom adopts the new five-year reviews under the new European Electronic Communications Code (EECC), combined with holistic/synchronised market reviews, Ofcom must consider the need for interim reviews and/or adjusting remedies mid-review (i.e. refreshing data and verifying the remedies in place are still appropriate). The need for adjustments mid-review are increased by both the longer review period as well as the high pace of change (much of which is unpredictable) over the next 10 years.
7. Most importantly, Ofcom must ensure that it conducts its market reviews by following the correct legal, economical and analytical approach, as set out in the relevant legislation and guidelines. While Ofcom has a strategy, this is not a relevant consideration when defining markets – defining economic markets to fit a certain ‘strategy’ is legally and economically wrong and increases the risk of successful challenge of Ofcom’s decisions, thereby increasing uncertainty. Ofcom’s strategy and objectives can be taken into account at the remedies stage, provided it is properly supported by analysis and evidence, and proper consultation. We are concerned that Ofcom may not follow due process while attempting to bend the regulatory framework to meet its own, internal vision of the ‘right’ regulatory direction
Product Market Definition
8. It is standard practice in market definition to define product markets before moving on to define geographic markets. This is because until it is understood what products customers are purchasing as substitutes for one another, there is no reasonable way to understand how far customers are willing to travel to strategy based on various product market definitions; there is no obvious reason why strategic goals should dictate product market definition. If Ofcom wishes to pursue this strategy, it can do so by choosing appropriate remedies to the SMP that it finds in objectively defined markets.
15. Ofcom should therefore conduct a proper, economic, product market definition exercise, starting from the narrowest possible focal markets and determine based on the hypothetical monopolist test whether they form separate product markets. This will allow for the correct market definition to be reached. There are a number of such focal markets which Ofcom should consider; in particular, Ofcom should assess whether multi-use networks constrain leased line specific networks (and vice versa) and whether coaxial cable based networks will impose a competitive constraint on FTTP and dedicated leased line networks (and vice versa). Ofcom should undertake this analysis in a transparent manner, carefully setting out the approach, which it proposes to adopt, and the evidence that it adduces in support of its conclusions.
16. There are a number of other important methodological issues which Ofcom should be cognisant of when defining product markets:
• Demand-side substitution analysis must be based on the competitive price (avoiding the cellophane fallacy), or markets can be defined too widely;
• Supply-side substitution analysis must only take account of capacity which can be switched without incurring meaningful costs, and with minimal time delay;
• The assessment must be undertaken on a modified greenfield approach basis, and in particular with no single pricing constraint— the hypothetical monopolist should be able to engage in price discrimination;
• The appropriate market definition may be asymmetric, and Ofcom should reflect this in its analysis.
Geographic Market Definition
17. As we discuss above, the precursor to the geographic analysis is a robust product market analysis, which identifies the product markets that will exist for the market review period 2021 – 2026.
18. Geographic market analysis is a process that follows on from the product market analysis and is used to identify the existence of variations in the levels of competition across the UK for these product markets. If there are geographic variations in the level of product competition between different geographic units, then a different approach to remedies may be justified, with it necessary to identify sub UK geographic markets for the product markets.
19. The period 2021 – 2026 is anticipated as being a period of FTTP network rollout. Today 4% of customers are able to access FTTP services. A range of prospective FTTP network builders have funding to build within the UK. Their plans are differentiated by their business / marketing strategies with the majority of companies proposing to serve a market niche and only one company proposing a wider scale network rollout in competition to Openreach at any scale. It is therefore premature to geographically segment the market.
20. The most significant market development for the period is most likely to be the upgrade by Openreach of the copper networks to G.fast and FTTP with an additional network in some geographies being installed alongside by alternative investors.
21. To understand how FTTP networks will be built in the UK it is necessary to consider a wider range of information that expands beyond just population density.(1) Ofcom should use a wider range of information and include the detail of the rollout plans from the prospective builders and prospective purchasers.
22. Understanding how the current retail market will respond is highly relevant to determining the level of expectant overbuild. We can see from other more Fibre developed EU countries that overbuild has tended to occur where that build has supported vertically integrated supply only.
23. Where a wholesale builder has entered the market there has been far less overbuild. Where a builder is a new entrant to the market and does not have an existing user base to covert to the service, first mover advantage into a geography is key.
24. As the number of locations with first mover advantage diminish the extent of commercially viable overbuild will reduce. To analysis geographies for longer-term competitive prospects it is essential to have a model which is capable of incorporating the current commercial reality of the UK market. Network builders will cherry pick geographies in which they can obtain the highest commercial success. That means being first to the geography and sweeping up the initial demand for full fibre services.
25. In practice in telecoms markets each premises is a separate economic market since there is no material demand-side or supply-side substitution. Rather than analysing each economic market, Ofcom proposes to agglomerate premises into categories by firstly combining premises into postcode sectors and secondly melding postcode sectors into categories that have certain competitive conditions. This introduces errors since a single geographic category will include many different economic markets that have very different competitive conditions and consequently the same remedies will be applied to areas with very different competitive conditions.
26. Whilst some such errors may be inevitable to ensure that the analysis is tractable, if Ofcom takes this approach it must recognise the existence of the variability of competitive conditions within each category. For instance:
• It should avoid creating wide categories since this will exacerbate errors and the harm from the errors, due to both under-regulation and over-regulation; and
• even in categories where there are BT+2 networks and BT’s share is below 40% it will be able to exploit market power by discriminating against pockets of premises in this category through raising prices and/or diminishing quality. Thus, in this case, an SMP finding would be appropriate in order to allow Ofcom to address this market power via a single pricing obligation.
27. A significant mistake Ofcom have made is to include in one category (category B) areas with very different competitive conditions (e.g. BT+2 actual and BT+0 actual) and apply the same remedies to all of these areas. This is inconsistent with the EC guidelines to only include in the same category areas “in which the conditions of competition are sufficiently homogeneous” and results in inappropriate remedies since the same remedy are imposed when different remedies are necessary. This exacerbates the errors introduced by its two-stage agglomeration approach. Furthermore, by adopting a wide category, Ofcom’s prejudices its approach to remedies since it is binding itself to imposing the same remedies across the whole area – if it adopted narrower categories, it could choose to impose different remedies or the same.
28. We consider that Ofcom is premature with respect to its vision of implementing a single market view across all product market types. We anticipate a retail product market analysis conducted for 2021 – 2026 would conclude that broadband (WLA) and leased lines (BCM) services fall into separate product markets having different service characteristics and different demand characteristics. The analysis is likely to find that a new product set for full fibre is emerging but its availability and take up is too limited to consider it as a significant part of the market.
29. Ofcom’s planned approach is better suited for consideration for the review 2026 – 2031. In the interim, Ofcom should encourage efficient and sustainable market entry by new FTTP networks. We consider this is best achieved by setting appropriate regulatory charges and appropriately monitoring the behaviour of Openreach and BT to ensure that network builders are given the opportunity to succeed.
Statistical approach/methodology to network presence assessments
30. We have a number of concerns with the proposed statistical approach to assessing where there is a competitive presence.
31. As discussed above, the ideal situation is for each premise to be considered as a separate market. However, we acknowledge there are challenges with this (although Ofcom is unclear on what data it does and does not hold). That said, the adoption of the 65% of premises with competing networks in a Postcode Sector is inappropriate, in that it will significantly over-estimate competition and is not sufficiently justified in this consultation. A justification based on the 65% rule having been used in other market reviews as far back as 2008, or incomplete reviews such as the BCMR 2018, is not sufficient.(2)
32. Ofcom say that a lower threshold is good to promote network competition and that it is better to over-identify areas as being competitive even when they are not.(3) As discussed above, the “strategy” of Ofcom to promote network competition should not be a relevant consideration in the geographic market definition process of a market review.
33. Furthermore, such an approach leads to a greater risk of customers who do not benefit from effective competition remaining that way, and even being worse off if Ofcom follows through this approach into remedies (as discussed below).
34. Given the potential harm to customers, we would urge Ofcom to err on the side of caution and have a premise-level analysis, or a much higher threshold to minimise this risk.
Rollout plans in “potentially competitive” areas
35. Ofcom also briefly discusses how it will take into account the different certainty levels of network builders plans (e.g. an intention, to Board-approved plans, to fully formed build plans, to works having started to install network/infrastructure).(4) It is clear that different weight should be applied to these in Ofcom’s decisions – something that Ofcom acknowledges. However, it is not clear how Ofcom will practically take account of different certainty into its assessment.
36. The concern is that the inclusion of areas in “potentially competitive” areas, and therefore lighter or no regulatory remedies, based on uncertain plans, which may never materialise is highly risky and could be damaging to customers and competition.
37. Ofcom needs to develop this further.
38. We are concerned that Ofcom is playing provider or investor in identifying potentially viable clusters of 20k premises, which could be economically viable.(5) Firstly, we do not agree that such analysis should form part of this review as no weight can be put in to such a designation. If such areas are not included in the plans of network builders, then there may be reasons/factors that Ofcom are unaware of.
39. Secondly, and as Ofcom note, the information that Ofcom has available to it is limited – with Ofcom suggesting a whole list of relevant considerations but admitting that it can only realistically examine size and population density. Without considering as many of the relevant factors as possible, the analysis will carry a greater level of inaccuracy, and lead to a risk of regulatory failure (e.g. over-deregulation leaving customers underserved by competition).
40. Thirdly, Ofcom indicates that the 65% rule will also apply to these 20k premise clusters. In that sense, Ofcom is content that only 13,000 premises need be covered to qualify as “potentially competitive”, and therefore quality for lighter or no regulatory remedies. This adds yet another layer of inaccuracy to Ofcom’s assumptions around competitive presence in these areas.
41. In summary on “viable clusters”, we think this approach is highly risky and based on very little evidence, and as such, we urge Ofcom to remove this kind of analysis from its network presence assessment, and rely on more confirmed evidence such as actual, existing network presence at premise level.
42. Ofcom’s current thinking on remedies appears to be as follows (based on discussions with Ofcom in December 2018):
• Category A areas – no SMP – no remedies
• Category B areas – no dark fibre (DFA), wholesale price cap above cost (MPF, GEA 40/10, Ethernet) or no price cap (GEA 80/20, FTTP)
• Category C areas – DFA, prices at cost
43. Our main concern is Category B, which Ofcom think, may be about 69% of UK. It includes a wide variety of areas:
• BT+2 networks (but SMP) and no additional networks expected;
• BT+1 networks, where additional networks may or may not be expected to greater or lesser degrees;
• BT+0 networks where Ofcom estimate there may be additional networks built to greater or lesser degrees.
44. Ofcom’s proposed remedies will unequivocally harm customers:
• Customers will pay excessive retail prices for many years before altnet FTTP investment occurs;
• Downstream competition will weaken, reducing the viability of altnet FTTP investment (which in turn will reduce Openreach’s incentive to invest in FTTP);
• Higher legacy product wholesale prices will reduce Openreach’s incentive to invest in FTTP (since incremental profits from investment are reduced);
• Dark Fibre Access delivers significant business customer benefits since it allows more innovation and competition than Ethernet, freeing retail providers from Openreach’s artificial bandwidth pricing gradient and their technology restrictions. It delivers almost as much innovation as network based competition yet avoids the significant duplication costs and scale economy losses of duplicating fibre, especially in locations where there is only likely to be one customer for that fibre;
• Notably Ofcom is proposing relaxed regulation in BT+1 areas (with no expected additional build) that have previously correctly been regulated based on cost-based prices.
45. In setting remedies, Ofcom must first ensure it has reached sound conclusions on product market definitions and have robust boundaries for any geographic segmentations for variations in competitive conditions.
46. Remedies on legacy products should only weaken when altnet FTTP investment actually occurs. This approach will reduce customer harm through unnecessarily high prices and will enhance FTTP investment.
47. UKCTA agree that there should be no price regulation of FTTP in this review period.
48. In Category A areas:
• Ofcom should only find no SMP when an areas is genuinely competitive e.g. not like CLA in BCMR where Ofcom claims no SMP despite Openreach’s market share being above 60%;
• Furthermore, Ofcom should recognise that, due to its use of postcode sectors and its agglomeration approach, in BT+2 areas, even where BT has market share below 40% BT will be able to exploit market power since in some areas it will be the only network and/or have much higher market share; and
• If there genuinely is no SMP, then Ofcom should consider if (and, if so, what) transitional measures are needed to protect customer interests.
49. In Category C areas, Ofcom should impose Dark Fibre Access and set prices at cost across all legacy products. Cost should exclude HON adjustments since no altnet investment likely.
1 As Ofcom indicates, it will in paragraphs 3.46 and 3.47.
2 Paragraph 3.24
3 Paragraph 3.23
4 Paragraphs 3.38 to 3.43
5 As discussed at paragraphs 3.46 to 3.50